CytoNorm: A Normalization Algorithm for Cytometry Data.

Area of Science:

• Immunology
• Computational Biology
• Biotechnology

Background:

• High-dimensional flow cytometry enables deep phenotyping of the human immune system at single-cell resolution.
• Data quality is challenged by sample preparation variability and measurement drift over time.
• Existing cross-sample normalization methods ignore single-cell data properties and risk removing biological signals.

Purpose of the Study:

• To develop CytoNorm, a novel normalization algorithm for high-dimensional flow cytometry data.
• To ensure internal consistency between clinical samples across different study batches.
• To address the limitations of current normalization techniques by preserving biologically relevant signals.

Main Methods:

• CytoNorm utilizes shared control samples across batches to learn batch-specific transformations.
• A clustering step identifies cellular distributions, followed by subset-specific quantile normalization using splines.
• Learned transformations are applied to clinical samples to correct for technical variations.

Main Results:

• CytoNorm demonstrated favorable performance compared to standard normalization procedures.
• Evaluation on customized and real-world clinical datasets confirmed the algorithm's effectiveness.
• The method successfully removed batch-specific variations while preserving cellular heterogeneity.

Conclusions:

• CytoNorm provides a robust approach for normalizing high-dimensional flow cytometry data.
• The algorithm ensures reliable and consistent analysis of immune cell populations across studies.
• CytoNorm is implemented as an R package, facilitating its application in clinical research.